dhi / fmdap

Pre- and post-processing for MIKE FM Data Assimilation

fmdap: MIKE FM Data Assimilation Processing tool

This python package contains pre- and post-processing tools for MIKE 21/3 Data Assimilation workflows.

Installation

> pip install https://github.com/DHI/fmdap/archive/main.zip

Requirements

• Python x64 >=3.6
• mikeio python package for reading and writing MIKE dfs files

Examples

Check out the examples in the notebooks folder.

Diagnostic outputs

• Diagnostic_output_overview.ipynb
• Diagnostic_output_plots.ipynb
• Diagnostic_output_skill.ipynb

Read settings from PFS file

• Pfs_file.ipynb

Estimate temporal and spatial correlation lenght scales

• Analyze_forcing_data.ipynb

fmdap's Issues

diagnostic output: iplot() is wrong for altimetry diagnostics

plotly probably cannot handle multiindex

Naming conventions and attributes

It seems that currently a lot of information is stored in the observation name and then at other places recovered from the string by splitting on space for example. This can go wrong as in this example. I think this is one use case of why we need to handle attributes in a more generic way in fmskill.

extract_DA_statistics detects whitespace in observation name as delimiter

log_helper.extract_DA_statistics detects whitespace in observation/measurement name as delimiter. Columns in returned dataframes are shifted as a consequence